motherboard pwr_on pins resistance?

The PC is turned off at night, not suspended or hibernated.

I always have current, redundant backups so no problem testing the HD.

No floppy so I'll try disconecting the HD or using a bootable CD and see
where it goes with another boot scenario.

I have replaced bad caps many times in other electronics and even have an
ESR meter. No visible signs on the MB.

MB is Gigabyte.

I appreciate that you took the time to respond and the advice. Thank you



"Paul" wrote in message
...
John Keiser wrote:
Paul,
If I may impose on your knowledge of PSUs, can you help me with an
earlier post?

I have a Acer Verition M460 [AMI BIOS R01-C3] with WindowsXP SP3.
I had the CMOS set to wake and boot every morning. Worked fine for
several years.
Then intermittent, now not at all.
In the morning the power light is on and the NIC light is sometimes on.
Screen has no info and the PC has stopped short of actually booting [so
Event Viewer has no info]. I simply hold the power button 3 seconds and
repower up. Always successful.
Time/date is accurate but I changed the CMOS battery and re-enabled the
RTC
alarm.
I also changed the PSU.
No joy.
Obviously not a major issue but I am curious, any idea why this is
happening?
Thank you.


When I google on Veriton M460, I'm seeing a higher than normal number of
problems there.

http://forums.techguy.org/virus-othe...e-related.html

http://forums.majorgeeks.com/showthread.php?p=1418793

Things you'd need to list in your query:

1) Was the system in S3 suspend to RAM, or S4 Hibernate, or just shutdown
from the menu ?
If you were starting up in the morning from S3, then the RAM could have
an issue.
(I have one machine here, that won't reliably start from S3 the next
day. The RAM
is good, so that's a motherboard problem. The motherboard can't retail
memory contents
overnight.)

If S4, then perhaps the drive isn't "becoming ready" within the timeout
interval.
If you're starting from a complete shutdown, that would be about the
same scenario
as Hibernate. Does the hard drive diagnostic that you can download for
it, pass or not ?
Seagate and Western Digital offer diagnostic programs. (Make sure
you've
burned the recovery media for your PC, in case the hard drive needs to
be
replaced at some point. Backups to an external hard drive would be nice
as well.)

2) Have you tested with some other boot scenario ? For example, set the
wake up time,
do a shutdown from the Windows menu. Then, plug a floppy diskette with
a copy of
memtest86+ on it, into the floppy drive. When the system starts the
next time, as
long as the floppy is first in the boot order, it'll boot from the
floppy. The purpose
of this kind of test, is to try to remove the hard drive from the
picture. Even better,
would be to unplug the hard drive data cable (so the system can't get
hung up,
while probing the hard drive).

http://www.memtest.org (scroll half way down, get the download,
use the program
to prepare a boot floppy. After prep, the
floppy cannot
be listed - there is no conventional file
system on it.)

Have you done a visual inspection of the motherboard recently ? Are there
any
bulging or leaking capacitors on the motherboard ?

http://www.badcaps.net/images/caps/kt7/kt3.html

If I could see some root cause, listed in the postings I can see for
Veriton M460,
that would give a better direction to look in. It sounds like it could be
a motherboard issue, but I suspect the machine may have been sold, with
more
than one motherboard type installed in it. (Which means, some versions of
the machine might have more problems than others, but the users wouldn't
list the motherboard details for us to know.)

You eliminated the power supply, so that's a start. They're a high runner,
in
terms of causing problems.

Failing power supplies, also give little hints about their health. For
example,
you may hear a muffled "sizzling" sound at startup. Or see a small puff of
smoke
go out the back of the machine at startup. Those are signs of failing
(leaking)
capacitors inside the supply. Another indicator I use for my personal
machines,
is when I notice a fixed speed 12V fan, is beginning to go "off pitch" on
its
fan sound. That can be an indicator of impending failure. It implies the
moment to moment drift of the 12V rail voltage, is larger than it used to
be. I used that to predict the impending failure of my very first supply.
That supply still "works" today, but the output voltage on the 12V rail
drops to 7V, with even a single 0.1 amp cooling fan connected to it.
So now, the supply is as "weak" as is physically possible. It puts out
less power, than a digital camera adapter :-) But technically, the power
supply still works, as under no load, all voltages are present and it's
cooling fan (internal one) still spins. It just can't take any load.
And internally, all the caps are flat, bright and shiny.

Not every failure condition has visual symptoms - but when offered a
visual symptom, go with it.

Paul

On Feb 1, 12:47*pm, Paul wrote:
mynick wrote:

thanks for great explanation
so in modern mobos the on switch grounds a pulled up line, straight to
southbridge
(Possibly there is a Schmitt trigger/inverter in between the two)

* * * * * * * +5VSB * * * * * * * * * * * * * * * * * * * * * * * +5VSB
* * * * * * * * | * * * * * * * * * * * * * * * * * * * * * * * * * *|
* * * * * * * Pullup * * * * * * * * * * * * * * * * * * * * * * * Pullup
* * * * * * * Resistor * * * * * * * * * * * * * * * * * * * * * * Resistor
* * * * * * * * | * * * * * * * * * * * * * * * * * * * PS_ON# * * * |
* * PWR * *X----+---- Motherboard logic ---- Open -------------------+- ...
* * * * * * * * * * * * * * * * * * * * * * *Collector *(to
* * GND * *X----+ * * * * * * * * * * * * * *Driver * * *ATX * * * * +
* * * * * * * * | * * * * * * * * * * * * * * * * * * * *supply) * * |
(Front * * * *GMD * * * * * * * * * * * * * * * * * * * * * * * * *GND
* Panel
* Switch)

Using an ohmmeter, between PWR and PS_ON#, doesn't tell you anything.
There is a silicon chip in the way.

* * Paul

Thanks
checked no shorts and it turned out empty bios batt prevented power
on,
however now there is no display with onboard or addon graphics
although rest seems ok(fans spinning,cpu warm,keyboard flashes)

Same result when booting to a CD: power light comes on but process proceeds
no further. Not a HD issue.

To clarify, the PC is set to awake from fully off using the RTC alarm, not
suspend or hibernate. This worked until a few weeks ago.

"John Keiser" wrote in message
news:4MadnfrN0cXZ59XQnZ2dnUVZ_sydnZ2d@powerusenet. com...
The PC is turned off at night, not suspended or hibernated.

I always have current, redundant backups so no problem testing the HD.

No floppy so I'll try disconecting the HD or using a bootable CD and see
where it goes with another boot scenario.

I have replaced bad caps many times in other electronics and even have an
ESR meter. No visible signs on the MB.

MB is Gigabyte.

I appreciate that you took the time to respond and the advice. Thank you



"Paul" wrote in message
...
John Keiser wrote:
Paul,
If I may impose on your knowledge of PSUs, can you help me with an
earlier post?

I have a Acer Verition M460 [AMI BIOS R01-C3] with WindowsXP SP3.
I had the CMOS set to wake and boot every morning. Worked fine for
several years.
Then intermittent, now not at all.
In the morning the power light is on and the NIC light is sometimes on.
Screen has no info and the PC has stopped short of actually booting [so
Event Viewer has no info]. I simply hold the power button 3 seconds and
repower up. Always successful.
Time/date is accurate but I changed the CMOS battery and re-enabled the
RTC
alarm.
I also changed the PSU.
No joy.
Obviously not a major issue but I am curious, any idea why this is
happening?
Thank you.


When I google on Veriton M460, I'm seeing a higher than normal number of
problems there.

http://forums.techguy.org/virus-othe...e-related.html

http://forums.majorgeeks.com/showthread.php?p=1418793

Things you'd need to list in your query:

1) Was the system in S3 suspend to RAM, or S4 Hibernate, or just shutdown
from the menu ?
If you were starting up in the morning from S3, then the RAM could
have an issue.
(I have one machine here, that won't reliably start from S3 the next
day. The RAM
is good, so that's a motherboard problem. The motherboard can't retail
memory contents
overnight.)

If S4, then perhaps the drive isn't "becoming ready" within the
timeout interval.
If you're starting from a complete shutdown, that would be about the
same scenario
as Hibernate. Does the hard drive diagnostic that you can download for
it, pass or not ?
Seagate and Western Digital offer diagnostic programs. (Make sure
you've
burned the recovery media for your PC, in case the hard drive needs to
be
replaced at some point. Backups to an external hard drive would be
nice as well.)

2) Have you tested with some other boot scenario ? For example, set the
wake up time,
do a shutdown from the Windows menu. Then, plug a floppy diskette with
a copy of
memtest86+ on it, into the floppy drive. When the system starts the
next time, as
long as the floppy is first in the boot order, it'll boot from the
floppy. The purpose
of this kind of test, is to try to remove the hard drive from the
picture. Even better,
would be to unplug the hard drive data cable (so the system can't get
hung up,
while probing the hard drive).

http://www.memtest.org (scroll half way down, get the download,
use the program
to prepare a boot floppy. After prep,
the floppy cannot
be listed - there is no conventional
file system on it.)

Have you done a visual inspection of the motherboard recently ? Are there
any
bulging or leaking capacitors on the motherboard ?

http://www.badcaps.net/images/caps/kt7/kt3.html

If I could see some root cause, listed in the postings I can see for
Veriton M460,
that would give a better direction to look in. It sounds like it could be
a motherboard issue, but I suspect the machine may have been sold, with
more
than one motherboard type installed in it. (Which means, some versions of
the machine might have more problems than others, but the users wouldn't
list the motherboard details for us to know.)

You eliminated the power supply, so that's a start. They're a high
runner, in
terms of causing problems.

Failing power supplies, also give little hints about their health. For
example,
you may hear a muffled "sizzling" sound at startup. Or see a small puff
of smoke
go out the back of the machine at startup. Those are signs of failing
(leaking)
capacitors inside the supply. Another indicator I use for my personal
machines,
is when I notice a fixed speed 12V fan, is beginning to go "off pitch" on
its
fan sound. That can be an indicator of impending failure. It implies the
moment to moment drift of the 12V rail voltage, is larger than it used to
be. I used that to predict the impending failure of my very first supply.
That supply still "works" today, but the output voltage on the 12V rail
drops to 7V, with even a single 0.1 amp cooling fan connected to it.
So now, the supply is as "weak" as is physically possible. It puts out
less power, than a digital camera adapter :-) But technically, the power
supply still works, as under no load, all voltages are present and it's
cooling fan (internal one) still spins. It just can't take any load.
And internally, all the caps are flat, bright and shiny.

Not every failure condition has visual symptoms - but when offered a
visual symptom, go with it.

Paul

John Keiser wrote:
Same result when booting to a CD: power light comes on but process proceeds
no further. Not a HD issue.

To clarify, the PC is set to awake from fully off using the RTC alarm, not
suspend or hibernate. This worked until a few weeks ago.


Have you added any hardware recently ? Like any USB devices ?
A USB card reader ?

A computer which crashes, before the BIOS can run any significant
number of instructions, won't beep (the computer speaker) at all.
The implication would be, that the default startup conditions
are no longer sufficient to allow the processor to start the
bootstrap process.

Some computers, they do a "double start" at startup. The BSEL
(bus select pins on the processor) request a particular FSB
setting (for processors that still have an FSB). The BIOS starts
to run, using nominal processor voltage and frequency. If the
motherboard is an enthusiast type, the BIOS can then set up
things according to any BIOS modifications and start the system
again. After this, you'd hear the single beep, which tells you
the machine has finished POST, and is about to start the OS boot
sequence.

If you're missing the single beep, and you're not getting an
error beep pattern, then the system is getting stuck early
on. Maybe not even one BIOS instruction has run in your case.

There is a second kind of failure for your situation, namely
the video display shows, but there is a flashing cursor in the
upper left hand corner. But if you had video, you probably would
have reported that here. ("Stuck with flashing cursor").

There are some simple things that can stop a board from
starting. The Power_Good signal coming from the power supply,
is a gating term in the reset chain. If the power supply
signaled that it wasn't at full voltage, by keeping Power_Good
de-asserted, that prevents the BIOS from booting. It's the
equivalent of pressing and holding the front panel reset button.
But since your system starts, after you've reset it, the problem
is not likely to be a power supply withholding Power_Good.

The motherboard adds terms to the Power_Good chain. Additional
status bits, can come from the onboard regulators. If an onboard
regulator withholds its own version of Power_Good, that stops
the startup process as well.

About all you can do at this point, is attempt a "clear CMOS" using
the Clear_RTC jumper. You do that with the power OFF and the computer
unplugged. The reason for that warning, is some older motherboards
will be damaged, if you use Clear_RTC while *any* power is present.
Usually, the user manual for the motherboard, warns about doing
a Clear CMOS with the system still powered. The purpose of
doing the Clear CMOS, followed by restoring all the BIOS settings
manually, is on the theory that some "phantom bit" in the 256 byte
CMOS RAM area, is causing the mis-programming of the motherboard
early in the BIOS startup/bootstrap sequence.

Other than that, its a good question as to what other persistent
storage on the motherboard, is remembering something it shouldn't.

Motherboards have "backfeed cut" circuits, intended to prevent
power from flowing in the wrong places, in the various shut off
states the motherboard can have. Such circuits can be fooled,
and not know the computer has been turned on, and then they prevent
the proper operation of the computer. (In one case, this was caused
by leakage current coming down the monitor cable - the computer would
start, if the user unplugged the monitor cable.) Backfeed circuit design
is kinda adhoc, so you're relying on the skills of the individual
motherboard designer, as to how well it works. Some aspects
of computer design are rather automated, or have design discipline
that ensures a high probability of success. But any time a designer
fiddles with analog power issues, there is room for error to crop
in, and the result is flaky behavior at startup. In some
cases, the behavior can even be temperature sensitive ("my
computer won't start, when the room is cold"). And for
all of those kinds of problems, I've got no answer. The
issue is complicated enough, that even with a schematic,
I'd be hard pressed to trace all those paths with a multimeter,
and determine what was going on.

Paul

mynick wrote:
On Feb 1, 12:47 pm, Paul wrote:
mynick wrote:

thanks for great explanation
so in modern mobos the on switch grounds a pulled up line, straight to
southbridge
(Possibly there is a Schmitt trigger/inverter in between the two)
+5VSB +5VSB
| |
Pullup Pullup
Resistor Resistor
| PS_ON# |
PWR X----+---- Motherboard logic ---- Open -------------------+- ...
Collector (to
GND X----+ Driver ATX +
| supply) |
(Front GMD GND
Panel
Switch)

Using an ohmmeter, between PWR and PS_ON#, doesn't tell you anything.
There is a silicon chip in the way.

Paul

Thanks
checked no shorts and it turned out empty bios batt prevented power
on,
however now there is no display with onboard or addon graphics
although rest seems ok(fans spinning,cpu warm,keyboard flashes)

Then you'd simplify the setup a bit, and retest.

At this point, are you hearing any BIOS beep pattern or not ?

Possible answers a

1) No beeps - implies processor can't run very much BIOS code.
RAM is bad in the low memory area. Processor 12V is unplugged.
BIOS code got erased by user accident.
2) One beep - implies processor has completed POST and is booting
the hard drive. Check for hard drive light flashing.
If the flashing "looks like Windows is booting normally",
it could be that only the video card output signal is
defective.
3) Multiple beeps - now we know the BIOS has detected some kind of
subsystem failure. Either video or RAM. If you
have two video devices, remove one and retest.
If you have two sticks of RAM, retest one stick
in place at a time.

When it comes to the display, your computer has multiple video connectors.
The output signal could be on a different connector, than the one
you're currently connected to. Alternately, the system is using the
built-in video, and wasn't able to detect the add-in video card.
Moving the monitor cable from connector to connector,
may be sufficient to sort that out.

You can remove the video card, and connect the monitor to the motherboard video.

If still no luck, power off the system, and use the Clear CMOS jumper.
This is on the theory, that the default BIOS settings include
enabling the built-in video. The system must be powered off, because
some motherboards get damaged, if you "Clear CMOS" with any power present.
I always unplug the computer, when doing this. After Clear CMOS,
if you can see a video screen, you'll need to re-enter the BIOS,
select the boot device and so on.

The fact that your keyboard flashes, implies you're not in (1) above.
So now it's a question of how many beeps you hear, is the hard
drive LED flashing a normal boot sequence, and so on. If could be
just a video issue of some sort.

Paul

On closer inspection, I do see a bulging cap near the cmos. Runs very hot
to touch so highly suspect.
Not sure if I can replace.


"Paul" wrote in message
...
John Keiser wrote:
Same result when booting to a CD: power light comes on but process
proceeds no further. Not a HD issue.

To clarify, the PC is set to awake from fully off using the RTC alarm,
not suspend or hibernate. This worked until a few weeks ago.


Have you added any hardware recently ? Like any USB devices ?
A USB card reader ?

A computer which crashes, before the BIOS can run any significant
number of instructions, won't beep (the computer speaker) at all.
The implication would be, that the default startup conditions
are no longer sufficient to allow the processor to start the
bootstrap process.

Some computers, they do a "double start" at startup. The BSEL
(bus select pins on the processor) request a particular FSB
setting (for processors that still have an FSB). The BIOS starts
to run, using nominal processor voltage and frequency. If the
motherboard is an enthusiast type, the BIOS can then set up
things according to any BIOS modifications and start the system
again. After this, you'd hear the single beep, which tells you
the machine has finished POST, and is about to start the OS boot
sequence.

If you're missing the single beep, and you're not getting an
error beep pattern, then the system is getting stuck early
on. Maybe not even one BIOS instruction has run in your case.

There is a second kind of failure for your situation, namely
the video display shows, but there is a flashing cursor in the
upper left hand corner. But if you had video, you probably would
have reported that here. ("Stuck with flashing cursor").

There are some simple things that can stop a board from
starting. The Power_Good signal coming from the power supply,
is a gating term in the reset chain. If the power supply
signaled that it wasn't at full voltage, by keeping Power_Good
de-asserted, that prevents the BIOS from booting. It's the
equivalent of pressing and holding the front panel reset button.
But since your system starts, after you've reset it, the problem
is not likely to be a power supply withholding Power_Good.

The motherboard adds terms to the Power_Good chain. Additional
status bits, can come from the onboard regulators. If an onboard
regulator withholds its own version of Power_Good, that stops
the startup process as well.

About all you can do at this point, is attempt a "clear CMOS" using
the Clear_RTC jumper. You do that with the power OFF and the computer
unplugged. The reason for that warning, is some older motherboards
will be damaged, if you use Clear_RTC while *any* power is present.
Usually, the user manual for the motherboard, warns about doing
a Clear CMOS with the system still powered. The purpose of
doing the Clear CMOS, followed by restoring all the BIOS settings
manually, is on the theory that some "phantom bit" in the 256 byte
CMOS RAM area, is causing the mis-programming of the motherboard
early in the BIOS startup/bootstrap sequence.

Other than that, its a good question as to what other persistent
storage on the motherboard, is remembering something it shouldn't.

Motherboards have "backfeed cut" circuits, intended to prevent
power from flowing in the wrong places, in the various shut off
states the motherboard can have. Such circuits can be fooled,
and not know the computer has been turned on, and then they prevent
the proper operation of the computer. (In one case, this was caused
by leakage current coming down the monitor cable - the computer would
start, if the user unplugged the monitor cable.) Backfeed circuit design
is kinda adhoc, so you're relying on the skills of the individual
motherboard designer, as to how well it works. Some aspects
of computer design are rather automated, or have design discipline
that ensures a high probability of success. But any time a designer
fiddles with analog power issues, there is room for error to crop
in, and the result is flaky behavior at startup. In some
cases, the behavior can even be temperature sensitive ("my
computer won't start, when the room is cold"). And for
all of those kinds of problems, I've got no answer. The
issue is complicated enough, that even with a schematic,
I'd be hard pressed to trace all those paths with a multimeter,
and determine what was going on.

Paul

On closer inspection, I do see a bulging cap near the
CMOS. Runs very hot to the touch, so highly suspect.
Not sure if I can replace.

Surround the area with absorbent material, then crush the capacitor. You
shold be able to tack-solder a replacement to the protruding leads.

I have practiced removing caps on old MB's and I find it not so easy.
I like this idea a lot and will give it a try.
Thanks for the suggestion.


"William Sommerwerck" wrote in message
...
On closer inspection, I do see a bulging cap near the
CMOS. Runs very hot to the touch, so highly suspect.
Not sure if I can replace.

Surround the area with absorbent material, then crush the capacitor. You
shold be able to tack-solder a replacement to the protruding leads.

John Keiser Inscribed thus:

I have practiced removing caps on old MB's and I find it not so easy.
I like this idea a lot and will give it a try.
Thanks for the suggestion.


Shush don't tell anyone, the secret is to preheat that area of the
board before you use the solder sucker... ;-)

Having said that some boards have so little clearance in the hole around
the pins that the crushing trick is the only way to free the pin so
that it will come out when heated from the underside, with the
soldering iron.

--
Best Regards:
Baron.

I modified the "crush" method: I snipped a test cap in half with wire
snips. The bottom half then pulls right off the internal connectors,
leaving two nice pegs for a tack solder.
Access could be an issue for this method, but very easy if there is room.



"Baron" wrote in message
...
John Keiser Inscribed thus:

I have practiced removing caps on old MB's and I find it not so easy.
I like this idea a lot and will give it a try.
Thanks for the suggestion.


Shush don't tell anyone, the secret is to preheat that area of the
board before you use the solder sucker... ;-)

Having said that some boards have so little clearance in the hole around
the pins that the crushing trick is the only way to free the pin so
that it will come out when heated from the underside, with the
soldering iron.

--
Best Regards:
Baron.